Russian Virtual Computer Museum. → An Essay on forming the Unified System of Electronic Computers (Part II)
An Essay on forming the Unified System of Electronic Computers (Part II)
by V.K. Levin, academician, member of the Russian Academy of Sciences
translated and edited by Alexander Nitussov
Read previous part: An Essay on forming the Unified System of Electronic Computers (Part I)
SRIDEC
In the second half of 1967 and during the whole 1968 all necessary practical works on the „Ryad“ computers were performed. Their purpose was to reach the speed levels 20-, 100-, and 500 thousand ops. Since the middle of 1968 those machines were named ES-1020, -1030, and -1050. They were manufactured in Minsk, Erevan and Moscow, correspondingly.
At the same period of time engineering teams were formed and many efforts were made to realize and promote the abovementioned governmental directive. When it was finally adopted (in December 1967), still bigger activity was needed to establish powerful scientific research centre for electronic computers in Moscow (SRIDEC), which should become the leading organisation for research and development of new ES Computers.
Initially, M.K. Sulim had the idea to form the SRIDEC by means of integrating already available Moscow institutes engaged in development of universal computers. However, administrations of those institutes did not support the idea and Sulim with his team could neither persuade nor force them to do that. The main reason was that they all were overloaded with own current works on the 2nd generation computers and their implementation in various concrete systems. Attention of their influential customers was also focused on those problems, thus, that time they demonstrated no interest to the 3rd generation computers.
Besides, those institutes were afraid to be dissolved in that „common pot“ of new centre, in case of their integration as M.K. Sulim had proposed. In spite of his own appointment as the deputy minister, in the spring of 1968, his efforts on creating new institute (SRIDEC) „from the very beginning“ also did not bring quick results.
Nevertheless, in autumn 1968 S.A. Krutovskikh (1928–1981), the director of the Institute of Electronic Machines (computers), changed his „autonomous“ position and accepted appointment as the director of SRIDEC and the chief designer of ES Computers.
I (the author) myself remained as his deputy (in fact, that was agreed with M.K. Sulim already in the beginning of 1968). During 1968 several groups of my (former) colleagues-specialists joined me in the SRIDEC. Before that, they were working on the pilot project of „Ryad“ computers (other name for ES Computers), however, despite their experience, total range of their specialties, qualification and professional profiles was not broad enough and, therefore, could not cover the whole spectrum of scientific and engineering activity necessary for such big work.
For that reason a big number of new young programmers – graduates from the Moscow State University and from the Moscow Institute for Physics and Engineering, were engaged in the SRIDEC work, and that became the decisive factor at adopting the software of „System 360“ on our grounds.
Professor Dr. M.R. Shura-Bura was invited to be the „guru“ of those young people. He was officially named „Scientific Supervisor of the ES Software Development Works“. He occupied positions at the Moscow State University and simultaneously at the Institute of Applied Mathematics (USSR Academy of Sciences); the work at the SRIDEC was his „volunteer duty“ (although, M.V. Keldysh – then the president of the Academy of Sciences also asked him about that). Dr. M.R. Shura-Bura was eventually awarded the State Premium.
Reorientation of the SRIDEC at such powerful organisation as the Institute of Electronic Computers with its rich experience in computers STRELA, M-20, M-220 (then, high-performance ones), with good manufacturing basis, developed technologies and well-developed infrastructure, quickly gave it sufficient impulse for active development. Works on the ES Computers gained needed rhythm. Of course, the governmental support was also very essential.
S.A. Krutovskikh displayed himself in general as a good leader and manager, although, his relations with M.K. Sulim suffered from repeating conflicts. However, in the beginning of 1970, Krutovskikh rather unexpectedly quit working, on health grounds, and was replaced with A.M. Larionov (1928–1995), also experienced collaborator of the Institute of Electronic Machinery. M.K. Sulim was also replaced, with famous academician V.S. Semenikhin (1918–1990) – an expert in military automated control systems.
We fruitfully worked together with Larionov until the end of 1975, when I became a director of another organisation, although my relations with the SRIDEC were not interrupted. In the end of 1977 Larionov, also on health grounds, changed his work for university teaching and V.V. Przhijalkovskiy was appointed as the SRIDEC director. In 1971 he was invited -as the chief engineer- from Minsk, where he was a recognized expert on computer engineering.
Also in the end of 1968, cooperation with the countries - members of the Counsil for Mutual Economic Assistance -CMEA: GDR, Bulgaria, Hungary, Poland and Czechoslovakia, was officially and organisationally established and its progress began.
That had both scientific-economical and political reasons. Romania sporadically joined some projects and, since 1973, Cuba also demonstrated interest in common development. The most useful and fruitful cooperation on engineering part of the ES Computers was with the East Germany (GDR), and with Bulgaria on development of memory devices on magnetic disсs and tapes; besides, such activity created international cooperation. Later, there was some mutual exchange between the countries with produced ES Computers means, facilities, etc.
The decision to produce ES Computers ready for both civil/universal and military implementations at the same time, was very important one. Many of us were accustomed for such production philosophy, sufficient experience had already been accumulated, so, there were no doubts that it would be the optimal way of engineering development. Manufacturing technologies implemented by military industry were focused on maintaining high level of reliability and quality. Those had already become adopted military manufacturing standards by that time, therefore, it was not difficult to extrapolate them onto civil production. Although, some expressed an opinion that such measures would make production process more complicated and increase the end price of new computers, in general there were no any serious objections. Later, efficiency of such approach fully proved to be true.
the standards for information coding and spacing on external carriers: punched cards and tapes, magnetic tapes and floppy disks, exchangeable disk stores and packs, etc. were adopted and implemented.
Many technical solutions incorporated in the Soviet computers of the 1st and 2nd generations were greatly different from those of the western engineering practices, which were also confirmed by decisions of international standartisation organisations – ISO and CCITT (International Telegraph and Telephone Consultative Committee). Adoption of the „System 360“ as the ES Computers prototype was essential stimulus for peripheral devices development to optimal level. The standards for information coding on them in their turn significantly influenced design of those devices.
Thus, implementation of internationally accepted standards, and development of their national analogues (in cooperation with GDR and Bulgaria), for such important memory devices as magnetic tapes and exchangeable discs was a significant step forward, because such parameters of our, previously produced, magnetic memory devices as speed, density and volume of recorded information were noticeably inferior.
Besides, there were also economical, political and organisational aspects of the ES Computers program progress depending on the general course of matters in the country, but it should be the subject for another narration.
Adoption of the system-prototype (System 360) made possible launching large-scale research program for ES Computers. Many other problems, which are commonly solved one after other (sequentially) in new developments, here were solved simultaneously.
Of course, alongside with the general progress many various obstacles hampered the work, among them political/economical impact of the Cold War, but that was so to say „inevitable“.
There were many difficulties with programming on real „System 360“ computers. Nevertheless, the funds were raised (the thing was an expensive one) and the dealers searched for the necessary software purchasing abroad.
In 1970 it was possible to send programmers to East Germany where there was one IBM computer suitable for programming research and experiments. The first „System 360“ -second-hand- licensed computer was put into operation in SRIDEC in 1971. Later there were also numerous direct contacts with the IBM, which by itself was interested in promotion of its brand (influence) and production in the East Europe, but that mutual understanding and agreements were nullified by countless bureaucratic obstacles on „higher levels“ (on all sides).
Simultaneously many leading specialists in Moscow, Minsk, Erevan and at the computer research and producing enterprises of (towns) Kazan and Penza were rather enthusiastic in that new development of the ES Computers.
For example, V.S. Antonov headed the works on ES-1050 and -later- ES-1060 at the SRIDEC. He gained good experience during development of the 1st and 2nd generation computers at the SDB-245. He was also distinguished veteran of the WW2 and respected scientist.
Those years many new specialists received solid experience and later became famous Soviet computer scientists and engineers. Some official and personal conflicts as well as other negative moments, which naturally also took place in such large scale process, eventually did not make essential negative influence on general situation.
Enormous technological and productional activity arose both at numerous institutes and industrial enterprises in connection with the ES Computers program. Soon that made computer producing industry one of the most advanced in the country and determined its development for the following decade. Among others the role of N.V. Gorshkov (1927-1995) should be mentioned. His merrits were awarded the highest Soviet title – the „Hero of Sosialist Labour“ (1983). Same could be said about the work of D.Ya. Berman, who created the leading experimental production enterprises, which timely satisfied needs of ES Computers development, on high quality level.
The Results
In summer of 1973, representative ES Computers presentation took place on the central exhibition site of Moscow. Production of all countries - the ES project members, was displayed there. It was not only a big engineering exhibition but also an international commercial event.
One more similar exhibition (actually a trade fair) took place in 1979; new ES computers were already compatible with the „System 360“ of IBM. Specialists named (or nicknamed) that level of computer development „three-and-half generation“ (program „Ryad-2“ in the USSR).
One should admit that then -and after- there were many critical discussions about „direct copying“ the IBM projects. However, that was far from the truth. The works on ES Computers undoubtedly permitted to make relatively quick progress and to raise these engineering and related fields to much higher level with progressive technologies, conceptions and generally bigger possibilities provided by the computers of the 3rdgeneration.
Software compatibility with the „System 360 and -370“, and resulting from it specifics of the following engineering and programming solutions, should be understood not as direct copying but as adoption of more progressive, high-quality standards, de-facto internationally recognized, which made possible broad usage of most modern and advanced software available on the world's market.
Such activity fully corresponded with legislative norms and regulations adopted (or could be adopted) it the country and society. There were also no patent problems (or conflicts) in the field of foreign trade. The ES computers reproduced only IBM solutions on the outer structure level („architecture“), what is not the object of patent rights. At the same time the concrete technical realisation was fully based on the own, original in many ways, research and design, and protected with the national patents (USSR author's/inventor's certificates).
One could also mention that similar approach to scientific, engineering and organisational solutions proved to be efficient and was/is widely used -both before and after described events- in the world's practice. Therefore, the situation with ES Computers was more or less similar with such events in other industrial branches.
The following facts could be considered as the examples of ES Computers positive effect.
When the ES program works just started, the software development (operating systems, translators, etc.) for the Soviet 2nd generation computers was only in the beginning stage. With the progress of programming for computers BESM-6, „Vesna“, „URAL“, etc. many features and single things were taken from the System 360 (ES Computers correspondingly) software and incorporated into it. Same could be said about terminology and general structure of the programming system.
The doubts about real possibilities of compatibility between the System 360 and the ES Computers evaporated, and later the strict correspondence between the Soviet and American computers was often demonstrated in various aspects, including official foreign trade. Also there were no problems with compatibility between different models of the ES computers themselves. There were many evidences of the positive effect of compatibility as the basic feature in the ES concepts. Unity and completeness of the ES Computers functional software eventually became a self-evident fact and the „conceptional conflicts“ and fears of possible negative side-effects connected with software compatibility were forgotten. Thus, one could claim realisation, confirmation and obtaining positive effect of implementation of the ES computers basic principles.
However, one couldn't say that there were no problems at all. Engineering realisation of the computers displayed signs of insufficient coordination in work of various enterprises – the project members. First of all, there was no perfect technical compliance, which was the characteristic of IBM performance. Reliability of computers as well as their operational characteristics leaved much to be desired and the project management failures were also not exceptions. The fields (and organisations) of their development, production and implementation were too much separated, coordination of their work was insufficient.
There were also unrealized technical possibilities; the „Systems 360, 370“ and, correspondingly, the ES Computers possessed very powerful complex of external data exchange means and promising input-output interface. Unfortunately possibilities of that complex were not fully used, even in the biggest computer centres. Partly that was caused with low technical level of some peripheral devices, however, foreign (better) devices also could be used and connected to those machines. Probably some functions of the ES computers were too advanced for the moment and there was not much practical (customers') need in them.
Well, anyway it's always scarcely possible to say in advance, what exactly should be done „to make things the best way“.
Already at development of the 1st and 2nd generation computers it was clear that more powerful machines should be joined into a net, to make them accessible for remote users (via terminals of that network). In 1966, the -future- academician A.P. Ershov (1931-1988) – one of the „brightest stars“ of the Soviet computing, proposed so-called „project AIST“ (network of automated stations/terminals). Some attempts were made to involve Ershov and his colleagues from the Siberian affiliate of the USSR Academy of Science into the ES Computers programs, but the long friendly diplomatic negotiations were the only „material“ result of them. However, his ideas and general approval played some positive role in ES computers progress.
Academician S.A. Lebedev (whom I personally honour as my teacher) in his detailed comments of 1967 on the „Ryad“ computers (on my request) paid special attention to the most powerful machines, and pointed up offering their services to users' terminals via networks. At the same time he admitted practical difficulties of technical maintaining of such communication channels. That was why, on his opinion, we needed more mini computers. He also did not see advantages in software compatibility of machines of different levels. He believed that each computer would solve problems of essentially different classes, that was why there wouldn't be need in transferring programs from one level to another one.
As the result the problem of networks caused increasing interest, so, many scientists, scientific teams and organisations enthusiastically „assaulted on the problem“; numerous projects and technical assignments were worked out and even a global national network (Integrated National Network of Computer Centres) were proposed. For that reason, already in the beginning of the ES Computers project work, there was special branch focused on remote data processing, analogs of the appropriate IBM prototypes were produced, network supporting software was released and the project specialists were often engaged in development of automated control systems.
Nevertheless, the Soviet communication technologies had great deal of traditional specific features in forming channels, in means for protection from interfering signals, in anti-jam facilities, in user terminals, etc. The resulting network systems were in fact a symbiosis of the Soviet and imported communication and information processing facilities, and became too cumbersome. Oppositely to them, data remote processing systems (data transmission networks) built on mini computers SM-4 in switching nodes and user terminals, and combined with more powerful ES Computers installed at the centres with increased volumes of computations, proved to be much more rational. American networks based on PDP-11 (SM-4 prototype), were also well-known (first of all military ARPANET).
However, combination of SM-4 and ES machines did not match the pattern of IBM software adaptation, which already became „traditional“, and the SRIDEC did not expose any interest, although some such hybrid systems were realized in some projects of other institute „Kwant“ (quantum).
In this connection one should admit (or even confess) that development and application of the ES Computers for systems and nets of multiple access lasted much too long. Only by 1985-1987 facilities with more or less significant number of video display terminals appeared able to maintain rather advanced interactive operation mode. However, that was already the time when the phenomenon of personal computers rushed into our life and the era of ES computers (in their classical form) came to the end.
Several episodes could illustrate development of ES computers, although some might find them too dull, others can say they are interesting. Thus, sharp arguing was caused with the beginning ES models, first of all ES-1050 performance evaluation, naturally, in comparison with foreign analogs and the Soviet BESM-6. In general, performance level of our computers was the highest among the European machines but often more or less inferior to the American ones. Nevertheless, due to more sophisticated programming technologies the Soviet specialists could efficiently solve real problems, what kept our „balance“ with the Americans.
In 1972-1974 we had to create „new philosophy“ on the advantages of the ES computers in comparison with BESM-6, which was the most popular computer of that time. The „junior“ machines ES-1020 and -1030 were to be compared correspondingly with computers „Minsk-22„ -32 and M-220. Later the performance evaluation methods became more „settled“ and the machines of ES series were compared only with others within that same system.
Of course, the work with new computers normally creates new problems, which are often underestimated in the beginning. All talks about easiness of operation with new machinery are but fairy tales for those who easily believe everything. Computing machinery is able to essentially extend our working possibilities, however, we have to pay for that with some additional efforts. Which are, by the way, not always acceptable for user. At that time development of Russian terminology (to replace the English computer lexicon „imported“ together with foreign scientific news) was in its beginning stage. At present it is advanced, of course.
One could remember many other things, and always a big percent of individual evaluations and personal opinions on significance of events is inevitable in such cases.
Thus, the great hopes were connected with the idea of microprogramming emulation by ES machines of all earlier computers (of the 2nd generation) on purpose of their available software transference and forming of a sort of unified software/data bank directly in machine codes.
There was also idea about emulation as the means of ES Computers specialization, what should increase their efficiency in certain application fields. However, all those efforts did not bring any essential results.
Another example; there were many attempts to begin regular serial production of applied program packages for common customers. Some -limited- results were really achieved. But it was only the wide distribution of IBM PCs, which efficiently brought „ready-made“ applied software to mass-production dimensions and turned it into real „market ware“ or a „life stile“.
Progress of computers with the highest performance (supercomputers) turned to be very complicated. Their development was often set aside from the mainstream, by the „current problems“ of serial computers mass-production.
The author was always convinced that the basic concepts of the ES Computers possessed all possibilities for increasing of their performance as much as the physical characteristics of the available microelectronic components permitted. Therefore, introduction of some other/new structural concepts was not necessary (unless those should be some very specific computing facilities).
Thus, development of ES-1060 (with planned performance 2-3 MIPS) and ES-1065 lasted too long and was hampered with many problems. That was only partly compensated with successful design and introduction of, replacing them, computers ES-1061 and ES-1066 (2 and 5 mill. ops correspondingly). The following works on the high-performance ES Computers did not bring practical results. All attempts of their serial production, beginning from ES-1050, were delayed or interrupted with unexpected re-orientation of the „manufacturing preferences“ of producing enterprises.
Unfortunately, the works on high-performance co-processors (with different variants of structural parallelism) were not given an appropriate industrial production basis and were not accomplished with sufficient development of applied programming oriented on specific features of parallel structures. Impressive success of other computers PS-2000, -2100, oppositely to the ES Computers, was a convincing evidence of scientific progress. Unfortunately, machines of those two types (PS and ES) were not compatible.
Main purpose of the author of the presented review consisted in demonstration of correctness and natural character of the ES system choice as perspective direction „in conditions of that time, place and way of things“. Consequences of any other of -then considered- variants (in case it would be accepted) could be negative or so.
Official approval and positive evaluations of the ES Computers, made several times, were „materialised“ in form of the state awards given to the scientific participants of the described events in 1983.
As the software compatibility with the IBM computers was always maintained, practical usage of the ES Computers made possible implementation of the most modern applied programs of that time and, therefore, realize functions of automatisation and information. If they were appropriately implemented in practice, is another question.
The judgments, evaluations and opinions given here have not only historical interest, they also influenced decision making on the next stages of computer science and engineering. Currently, experience of the ES Computers development is implemented in the modern „main frame“ projects. Software compatibility conceptions are largely interpreted with „programming portability“ (transportability, continuity) demands. These are implemented (in own way) in new generations of PCs and working stations.
Compatibility can be a natural trait both in various servers and supercomputers, because of the scalability of systems (on number of processor units, volume of storage, etc.) and also with uniformity of programming means (at least on higher level). One can notice that currently there is a number of works on the mentioned directions. Actual needs of the Russian customers and users are satisfied, there is also a vision of the ways of progress in the nearest future [1, 2].
Here the author didn't attempt to provide balanced characteristics of all aspects of the ES Computers history. These are only documents and personal direct impressions and views that, taken together, present only scientific and organisational level of the problem, but not the common views, opinions or social and economical generalisations. More details could be given in the narration but the resulting conclusions would remain the same.
It should be also mentioned that the speculations limited only with the ES Computers subject can not demonstrate the whole landscape of the computer science, engineering and production in the country. Of course, there were many other popular and „closed“ works on computers and computations, nevertheless, many of them were directly or indirectly influenced by the ES progress.
A serious new program – new line of computers, an alternative to the ES line was considered in the 1983-1985, in connection with making new plans of computer production development in 1986-1995.
Same as in other cases here in this field there was real entangling of objective and individual circumstances, what predetermined decision making, besides, that is not the subject of our narration. The beginning „perestroika“ processes in the USSR of that time basically changed general orientation, evaluations and opinions.
The author expresses his gratitude to all who were engaged in the ES Computers development and is really sorry for just having briefly mentioned only a few of their names, as there was no possibility to properly display contribution and role of every participant of that impressive scientific and engineering undertaking.
The original article was published 24.02.2004
Sources:
- Levin V.K. Radioelectronics and Computers. Radiotekhnika (publ.) 1995 №4-5 pp. 137-141.
- Levin V.K. High performance Computing Systems for Solving Scientific and Industrial Problems in Russia. In: Information Technologies and Computing Systems, 2003, № 4.
Published in museum, 14.02.2016
